In E-UTRAN (Evolved-Universal Terrestrial Radio Access Network, also called 3GPP) Orthogonal Frequency Division Multiple Access (OFDMA) technology is used in the downlink. OFDM is a modulation scheme in which the data to be transmitted is split into several sub-streams, where each sub-stream is modulated on a separate sub-carrier. Hence, in OFDMA based systems, the available bandwidth is sub-divided into several sub-channels called resource blocks (RB) or units, in both uplink and downlink. A resource block is defined in both time and frequency. According to the current assumptions, used herein, a resource block size is 180 KHz and 0.5 ms (time slot) in the frequency and time domains, respectively. The resource block size in the time domain, here 0.5 ms, is often called time slot. One or more resource blocks are allocated to a User Equipment (UE) for data transmission. The transmission time interval (TTI) comprises 2 time slots, which correspond to a sub-frame of 1 ms length in time. The radio frame is 10 ms long i.e. comprising of 10 sub-frames. The overall uplink and downlink cell transmission bandwidth can be as large as 20 MHz; other typical bandwidths are 1.4, 3, 5, 10 and 15 MHz. In the case of 20 MHz bandwidth up to 100 resource blocks (RB) containing data and control signalling can be transmitted by the UE in the uplink or by the network, e.g., a base station, in the downlink. The UE can be allocated a sub-set of the resource blocks for reception and transmission of data and control signalling.
Downlink Neighbour Cell Measurements for Mobility
In WCDMA the following three downlink neighbour cell measurement quantities are specified primarily for mobility purposes:                1. Common PIlot CHannel (CPICH) Received Signal Code Power (RSCP), the received power on one code after de-spreading measured on the pilot bits of the CPICH. The reference point for the RSCP is the antenna connector at the UE.        2. CPICH Ec/No; CPICH Ec/No=CPICH RSCP/carrier RSSI, where RSSI=Received Signal Strength Indicator. CPICH Ec/No can be described as the received energy per chip divided by the power density in the band. Measurement is suitably performed on the CPICH. The reference point for CPICH Ec/No is the antenna connector at the UE.        3. UTRA carrier RSSI, can be described as the wide-band received power within the relevant channel bandwidth. Measurement is suitably performed on a UTRAN downlink carrier. The reference point for the RSSI is the antenna connector at the UE.        
Reference [1] describes downlink neighbour cell measurements for WCDMA more in detail.
The RSCP is measured by the UE on cell level basis on the common pilot channel (CPICH). The UTRA carrier RSSI is measured over the entire carrier. The UTRA carrier RSSI is the total received power and noise from all cells (including serving cells) on the same carrier. The above CPICH measurements are quantities that are often used for the mobility decisions.
In E-UTRAN the following three downlink neighbour cell measurement quantities are specified also primarily for mobility purposes:                i. Reference symbol received power (RSRP)        ii. Reference symbol received quality (RSRQ): RSRQ=RSRP/carrier RSSI        iii. E-UTRA carrier RSSI        
Reference [2] describes downlink neighbour cell measurements for E-UTRAN more in detail.
The RSRP or RSRP part in RSRQ is solely measured by the UE on cell level basis on reference symbols. As in the case of WCDMA, the E-UTRA carrier RSSI is measured over the entire carrier. It is also the total received power and noise from all cells (including serving cells) on the same carrier. The two RS based measurement quantities (i. and ii.) are often used for the mobility decisions.
The neighbour cell measurements are averaged over a long time period, in the order of 200 ms or even longer, to filter out the effect of small scale fading.
There is also a requirement on the UE to measure and report the neighbour cell measurements, e.g. of RSRP and/or RSRQ in E-UTRAN, from a certain minimum number of cells. In both WCDMA and E-UTRAN this number is often 8 cells (comprising one serving and seven neighbour cells) on the serving carrier frequency. The serving carrier frequency is commonly called intra-frequency. Hence, the expression “neighbour cell” includes both the serving cell of an UE and the neighbour cells of this serving cell.
Sampling of Neighbour Cell Measurements
The overall neighbour cell measurement quantity results comprises non-coherent averaging of 2 or more basic non-coherent averaged samples. An example of RSRP measurement averaging in E-UTRAN is shown in FIG. 1. The figure illustrates that the UE obtains the overall measurement quantity result by collecting four non-coherent averaged samples or snapshots, each of 3 ms length in this example, during the physical layer measurement period, e.g. 200 ms. Every coherent averaged sample is 1 ms long. In this example a 3 ms non-coherent sample comprises 3 consecutive coherent samples. The measurement accuracy of the neighbour cell measurement quantity, e.g. RSRP or RSRQ, is specified over the physical layer measurement period. It should be noted that the sampling rate is UE implementation specific. Therefore in another implementation a UE may use only 3 snap shots over a 200 ms interval or measurement period. Regardless of the sampling rate, it is important that the measured quantity fulfils the performance requirements in terms of the specified measurement accuracy.
In case of RSRQ both RSRP, numerator, and carrier RSSI, denominator, should be sampled at the same time or instant to follow similar fading profiles on both components.
Mobility Scenarios
There are basically, or at least, two kinds of mobility:                a. Idle mode mobility: cell reselection        b. Connected mode mobility: handover        
The cell reselection is mainly a UE autonomous function without any direct intervention of the network. But to some extent the behaviour of the UE in this mobility scenario could still be controlled by some broadcasted system parameters and performance specification.
The handover is on the other hand often fully controlled by the network through explicit UE specific commands and by performance specification.
In both idle and connected modes the mobility decisions are mainly based on the same kind of downlink neighbour cell measurements, which were discussed previously.
Both WCDMA and E-UTRAN are frequency reuse-1 systems. This means that the geographically closest or physically adjacent neighbour cells operate on the same carrier frequency. An operator may also deploy multiple frequency layers within the same coverage area. Therefore, idle mode and connected mode mobility in both WCDMA and E-UTRAN could be broadly classified into three main categories:                Intra-frequency mobility (idle and connected modes)        Inter-frequency mobility (idle and connected modes)        Inter-RAT mobility (idle and connected modes)        
In intra-frequency mobility the UE moves between the cells belonging to the same carrier frequency. This is an important, maybe even the most important, mobility scenario since it involves less cost in terms of delay due. This mobility scenario involves shorter delay since UE measurements are not done during the measurement gaps. Secondly most handovers and cell reselections are carried out between the cells operating over the same carrier frequency. In addition, an operator would have at least one carrier at its disposal that it would like to be efficiently utilized.
In inter-frequency mobility the UE moves between cells belonging to different carrier frequencies but of the same access technology. This could be considered as a less important mobility scenario than intra-frequency mobility. This is because handover and cell reselection between cells belonging to different carriers are carried out when no suitable cell is available on the serving carrier frequency. Furthermore, UE measurements for inter-frequency mobility are done in gaps. This increases measurement delay and consequently involves longer handover delay compared to that in case of intra-frequency scenario.
In inter-RAT mobility the UE moves between cells that belong to different access technologies such as between WCDMA and GSM or vice versa. This scenario is particularly important in case an operator does not have full coverage of all the supported RATs in its network. During an initial deployment an operator may have limited coverage of the newly deployed technology. Thus inter-RAT handover would ensure ubiquitous service to the users even if all RATs don't have full coverage. Furthermore, an operator may optimize different RATs for different services e.g. GSM for speed, UTRAN for packet data and E-UTRAN for both speech and packet data. Thus if UE switches between speech and packet data or requires both type of services at the same time then if necessary, an inter-RAT handover can be used by the operator to select the most appropriate technology for offering the requested service to the prospective subscriber.
Objectives of Quality Measurements
As indicated above, CPICH Ec/No and RSRQ are so-called neighbour cell quality measurement quantities used in WCDMA and E-UTRAN respectively.
In general the quality measurement (Qrx) can be expressed as follows:
                              Q          rx                =                              P            rx                                I            +                          N              o                                                          (        1        )            
Where, Prx is the received power of the pilot or reference signal or channel, i.e. signal strength part, I is the interference and No is the noise. Depending upon the type of quality measurement the component I can be interference on the pilot channel or the total interference on the entire carrier or simply inter-cell interference plus noise. In current quality measurements in WCDMA and in E-UTRAN the interference measurement constitutes the entire interference on the carrier i.e. from the serving and all non serving cells. In reality the noise and the interference within the same measurement bandwidth cannot be separated. This means that the interference measured by the UE would incorporate both the actual interference and the noise i.e. the measured part is the entire denominator (I+NO) in (1).
The goal of the neighbour cell quality measurement is to estimate and predict the long term downlink quality that can be experienced by the UE in a particular cell. It should indeed indicate the signal quality or throughput that the UE will achieve in a cell. This prediction enables the UE and the network to choose the most appropriate cell when performing cell reselection and handovers, respectively. In E-UTRAN any set of resource blocks (i.e. part of the cell bandwidth) can be assigned to the UE for transmission. Therefore the quality measurement should capture the overall long term average quality over the entire bandwidth or at least over the largest possible portion of the bandwidth. This is in contrast with E-UTRAN CQI measurement, which typically depicts short term quality of possibly a sub-set of the resource blocks from the serving cell.
Problems with Existing Solutions
As noted above, the quality measurements include the total interference on the entire carrier in their denominator e.g. RSRQ=RSRP/carrier RSSI. This means that the quality measurement also includes a contribution from the serving-cell signal. Especially in a OFDM based system like E-UTRAN the serving-cell signal introduces negligible intra-cell interference due to good orthogonality between the sub-carriers across the cell bandwidth. In order to correctly track the cell quality the contribution from the serving-cell signal should hence be excluded from the interference measurement part of the neighbour cell quality measurement.
Furthermore, the statistical characteristics of the inter-cell interference may be significantly different, depending on whether the inter-cell interference originates from:                1. reference symbols from neighbouring cells        2. data signaling from neighbouring cells        3. control signaling from neighbouring cells        
Each of these three categories can have different transmission power and spatial characteristics.
For accurate neighbour cell quality measurement or estimation the UE must have good, statistics, or suitably should obtain statistics or statistical characteristics, of the inter-cell interference that is hitting or affecting the resource elements (RE) in the data channel, which is a mixture of the three categories mentioned above, here referred to as I_d. One may also use the expression, the inter-cell interference from the resource elements (RE) in neighbor cells, where the REs suitably should belong only to the data channel. But the REs may also be a mixture of one or more of the three categories of REs; data signaling, control signaling and reference symbol containing REs.
The statistical characteristics of the inter-cell interference is here referred to as I_d.
In reality, at least often, the inter-cell interference that is hitting or affecting the resource elements (RE) in the data channel is a mixture of the three categories 1, 2 and 3 mentioned above.
Ultimately or suitably, this interference statistics should be measured, or measured and calculated, on the data channel itself. However, this measurement is limited to resource entities, i.e. or e.g. time-frequency resource elements, that contain data scheduled to the particular user or UE doing the measurement since there is only or mainly a good chance of removing the contribution from the serving-cell signal for resources where the UE doing the measurement is scheduled. One may also say, for resources over which the UE is receiving data sent by these resource entities. The limited number of interference samples can significantly penalize the accuracy of the statistics estimate, or of the measured or estimated interference statistics. Moreover, in multiuser-MIMO (MIMO=Multiple Input Multiple Output), i.e. or e.g. spatial division multiple access, systems, several users may be assigned the same data resource elements (RE), which in effect prohibits the UE to separate the inter-cell interference from the intra-cell interference, if the measurement is performed on data REs.
Alternatively, the interference measurement can be performed on REs containing reference symbols (RS). However, the statistics, or the statistical characteristics, e.g. average interference, of the interference measurement on the neighbour cell RS, corresponding to I_RS, may have significantly different statistics than the interference on the data channel, or the control channel. It may also be the case that the statistical characteristics of the interference measurement, I_RS, on the neighbour cell RS is different, or significantly different, than the statistical characteristics of the interference measurement on the data channel, or of the interference measurement on the control channel. The interference measurement on the neighbour cell RS gives or yields the interference from the neighbour cell RS. There is a limited set of RS:s and in particular for MIMO, where the position holding a RS on one antenna is empty for a neighbouring antenna. Alternatively, one may say that for MIMO the time-frequency resources, i.e. the resource elements, containing the RS on different antennas are different. Therefore the interference hitting a RS will to a larger extent, or mainly, come from, or be contributed by, the RSs of the neighbouring cells. For example in lightly loaded systems, I_RS may be significantly different, typically or often substantially larger, than I_d, because possibly data is not allocated to all resource blocks (RB) in the neighbouring cells. The statistics of the measured interference term may therefore deviate significantly from the interference that hits the data channel. The RS grid for a RB in case of 1, 2 and 4 transmit antennas is illustrated in FIGS. 2a-2c4 Between cells, the RS grid may be shifted in the frequency domain. This is because the standard allows the possibility of configuring three possible shifts in frequency domain to allow the randomization of the interference. The frequency shift used in a cell is mapped on to primary synchronization sequence (PSS). Therefore three unique PSS are possible. The frequency shift is detected by the UE during the cell synchronization phase, which requires the detection of PSS. One RS grid may often span over a time slot, 0.5 ms, or a sub-frame, 1 ms, in the time domain and over the entire cell bandwidth (BW) in the frequency domain. In the frequency domain that is over multiple RBs, e.g. 50 RBs in a cell with 10 MHz BW or 100 RB:s in a cell with 20 MHz BW and so forth. Reference sign 202 indicates a resource element, which may be identified by an index (k,l) where 1 ranges from 0-6 and k from 0 to 12.
For 2 transmitting antennas only three frequency shifts for common RSs exists. This will lead to that not all data interference can be measured. Furthermore, the first three OFDM symbols might see control channel interference instead of data interference. Since control signaling may be differently power controlled than the data signalling, the interference estimate obtained on these RS s may not reflect the interference present when data is transmitted. If common RSs in the later part of a sub-frame is removed, for example because dedicated RSs are inserted instead, it might be necessary to measure interference on data REs.